The central hypothesis of this proposal is that some articulations covary in distinctive features and segments because their acoustic effects are integrated in perception, rather than because those articulations are yoked to each other physiologically. this hypothesis will be tested using two paradigms designed to assess perceptual integration of orthogonal stimulus dimensions: the Garner (1974) paradigm to assess integration among a variety of articulatory covariates in vowel height and [voice] contrasts, and the Nearey (1990) paradigm to assess integration of redundant with distinctive features in labial consonants. This set of experiments focuses on selected contrasts, but investigates each of them comprehensively. If integration can be demonstrated among the acoustic effects of the covarying articulations in these distinctive features and segments, then this aspect of the listener's response to speech signals contributes substantially to reducing the degrees of articulatory freedom allowed the speaker, by selecting those articulatory aggregates that best serve the integrative perceptual mechanisms in the listener. Among other things, this result implies that the acquisition or remediation of articulatory control would be best served by a demanding rather than tolerant listener.